What is better for gamming, 3400+ ClawHammer or 3400+ Newcastle? Or are they about the same in performance? I do see that the Newcastle is at 2.4 and the Claw is 2.2 but the Claw has 1meg of cache. Thank you for your time reading this thread. Take care all.
They are our, fathers, mothers, brothers, sisters, sons, daughters, friends, neighbors. They are OUR TROOPS! Please support them.
i think A64 3400+(2.4)512k newcastle is better, than A64 3400+(2.2)1mb. However clock for clock the clawhammer is much better. for example, the A64 3700+(2.4) clawhammer has higher PR rating than the A64 3400+(2.4)newcaslte.
I don't think you can make that a general statement. It depends on how games/apps are programmed. If an app is heavily depending on cache then extra cash will pay off better.
actually, you can . because overall, the A64 architecture really doesn't benefit from extra cache like the P4s do. Just like dual channel didn't really significantly boost real world performance on A64s, such as office productivity, or general usage applications. Dual channel and Cache are really for more beneficial on P4 netburst architecture because AMD64 uses hypertransport and onboard cache becomes less effecient and important. A64's have less latency than P4s and higher IPC , so cache doesnt really effect performance as much.
A64 3400+ Newcastle
1GB PC 4000 Kingston HyperX
Asus K8V basic Bios 1004
PNY Geforce 6800 GT 256MB DDR3
63,524 Aquamarks <P ID="edit"><FONT SIZE=-1><EM>Edited by Kanavit on 09/10/04 11:07 AM.</EM></FONT></P>
Hi Kanavit. I've been through this before, but am hoping you will listen now. The P4s need fast memory access, and a lot of it, because of prefetch. Mostly what that means, is that the Intel chips have to do extra dips at the memory well. See, prefetch is an attempt to prevent errors, which are a real concern on netburst.
Intel chips will have to do more memory calls than an Amd chip would, for the same process.
AMD64 uses hypertransport and onboard cache becomes less effecient and important.
Actually HTT feeds more info to the chip, so it is better able to make use of cache. Since the cache is not filled with prefetch bs, it is used more eficiently, though the amount of cache required is less.